View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Bern Open Repository and Information System (BORIS)

Paternity of Subordinates Raises Cooperative Effort in

Rick Bruintjes1,2*, Danielle Bonfils1, Dik Heg1,3, Michael Taborsky1 1 Department of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, Hinterkappelen, Switzerland, 2 School of Biological Sciences, University of Bristol, Bristol, United Kingdom, 3 Institute of Social and Preventive Medicine (ISPM), University of Bern, Bern, Switzerland

Abstract

Background: In cooperative breeders, subordinates generally help a dominant breeding pair to raise offspring. Parentage studies have shown that in several species subordinates can participate in reproduction. This suggests an important role of direct fitness benefits for cooperation, particularly where groups contain unrelated subordinates. In this situation parentage should influence levels of cooperation. Here we combine parentage analyses and detailed behavioural observations in the field to study whether in the highly social pulcher subordinates participate in reproduction and if so, whether and how this affects their cooperative care, controlling for the effect of kinship.

Methodology/Principal Findings: We show that: (i) male subordinates gained paternity in 27.8% of all clutches and (ii) if they participated in reproduction, they sired on average 11.8% of young. Subordinate males sharing in reproduction showed more defence against experimentally presented egg predators compared to subordinates not participating in reproduction, and they tended to stay closer to the breeding shelter. No effects of relatedness between subordinates and dominants (to mid-parent, dominant female or dominant male) were detected on parentage and on helping behaviour.

Conclusions/Significance: This is the first evidence in a cooperatively breeding fish species that the helping effort of male subordinates may depend on obtained paternity, which stresses the need to consider direct fitness benefits in evolutionary studies of helping behaviour.

Citation: Bruintjes R, Bonfils D, Heg D, Taborsky M (2011) Paternity of Subordinates Raises Cooperative Effort in Cichlids. PLoS ONE 6(10): e25673. doi:10.1371/ journal.pone.0025673 Editor: Nicolas Chaline, Universite´ Paris 13, France Received March 17, 2011; Accepted September 9, 2011; Published October 12, 2011 Copyright: ß 2011 Bruintjes et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The project was supported by the Swiss National Science Foundation (SNSF grants 30100A0-105626, 31003A_118464 and 31003A_122511 to Dr. Taborsky). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]

Introduction participating in reproduction [13–15] (see [16] for a review of

| downloaded: 8.5.2016 hypotheses). Cooperative breeding, where subordinates help dominants to The importance of current direct fitness benefits to subordinates raise offspring, is rather widespread in vertebrates [1–5]. This obtained through parentage acquisition for the decision to help has helping behaviour has puzzled evolutionary biologists for a long been questioned, partly due to assumed monopolization of time, as it is costly and often does not generate obvious fitness reproduction by dominants [3,17]. However, studies of several benefits to subordinates [6–8]. Several hypotheses have been cooperatively breeding vertebrates have found intra-group repro- proposed to explain the regulation of helping behaviour in ductive participation of subordinates (e.g. in fish: [18]; birds: [19] cooperative breeders. First, the kin selection hypothesis predicts and mammals: [20]), which suggests a potential for direct fitness that subordinates should raise their level of help with increasing benefits of subordinates due to care of own offspring. In relatedness to recipients to acquire indirect fitness benefits, polyandrous birds with cooperative care, for example, a positive contingent on the relationship between benefits to recipients, costs association has been found between receiving a share in mating - to subordinates, and the relatedness between them [7]. If but not necessarily in parentage - and subordinate investment subordinates and dominants are only distantly related or [21,22], which highlights the potential importance of direct fitness unrelated, several mutually non-exclusive alternative hypotheses benefits for the brood care effort of subordinate males. In attempt to explain helping behaviour. (i) The prestige hypothesis cooperative breeders, a positive association between subordinate proposes that subordinates may help dominants to signal their relatedness and helping effort has been observed in several cases genetic quality to potential future partners [9]. (ii) The group (e.g. [13,23,24]), whereas in others this did not hold [25,26] or the augmentation hypothesis proposes that cooperative care is selected results were mixed [14,27,28]. As yet, a positive relationship https://doi.org/10.7892/boris.7339 by beneficial effects of group size [10]. (iii) The pay-to-stay between subordinates’ parentage and their helping effort has been hypothesis proposes that subordinate helping serves as payment rarely found in cooperative breeders in the wild (but see [22,29] for being allowed to stay in the group [11,12]. Finally (iv), helping for support in birds and mammals). Nevertheless, in cooperatively subordinates may accrue current direct fitness benefits by breeding fish, one experimental laboratory study showed that source:

PLoS ONE | www.plosone.org 1 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation female subordinates unrelated to the dominants performed more the fish use distinct stone patches for shelter and breeding, created alloparental brood care when they acquired a share in reproduc- by digging away sand [45,48]. Predation risk is of key importance tion (i.e., when they were allowed to produce own clutches [30]). for group living in N. pulcher, since subordinates are protected by In the cooperatively breeding cichlid fish Neolamprologus pulcher, larger group members [34]. Subordinate relatedness towards subordinates are often distantly related or unrelated to the newborn fry (i.e., beneficiaries) diminishes with age, due to high dominants, due to the high philopatry of subordinates and high turn-over rates of dominants and high philopatry of subordinates turn-over rates of dominants, which is most pronounced in males [31,32]. As a consequence, subordinates often help to raise non- [31,32]. The main benefit of subordinates to stay in a territory of kin broods [33,49], which increases the productivity of dominants dominant breeders is the protection gained against predators, [33,50] and lowers their work load [45,51]. This service of which is provided by the large group members [33,34]. By subordinates is provided as payment or ‘rent’ for being tolerated participating in reproduction, subordinates face a threat of and protected in the dominants’ territory [36,40,42,51–53]. eviction [35], which may be detrimental due to the high mortality Furthermore, group stability was shown to increase with group risk outside of territories [36]. In the laboratory, male subordinates size [44]. Recently it has been shown that N. pulcher is able to unrelated to the breeding pair were found to participate in recognize relatives [54]. The effect of relatedness between reproduction [31,35,37,38]. However, their reproductive role in subordinates and dominants on helping effort has revealed mixed nature was questioned because subordinates have smaller gonads results in the field and in the laboratory, where relatedness than breeders [17]. At present, data on subordinate parentage between dominants and subordinates was negatively associated under natural conditions are lacking [39], and potential effects on with helping effort in one study [28], and not associated with subordinate helping effort are unknown. helping levels in another [26]. In this study we combine parentage analysis and detailed behavioural observations to investigate if subordinates participate Set-up and sampling in reproduction in the field and if so, whether and how this affects Group territories in our study population were mapped and subordinate helping behaviour. Due to low relatedness between marked with numbered stones. Experimental units were created dominants and subordinates, mature male subordinates can by haphazardly selecting two adjacent N. pulcher group territories accrue only minor indirect fitness benefits by helping, which with groups composed of at least one breeding pair, one large might provide incentives to acquire current direct fitness benefits (.37.5 mm SL) and one small subordinate (25–37.5 mm SL). We through parentage acquisition. Therefore, we predicted that male used groups with differently sized subordinates, because of subordinates should participate in reproduction in the field and demonstrated size-dependent responses to demand, and size- that their helping effort should be contingent on paternity dependent task specialization [42,52]. Small subordinates defend acquisition. Finally, we assessed whether relatedness between more against egg predators coming close to the breeding shelter subordinates and dominants might affect parentage acquisition than large subordinates, whereas the latter were shown to spend and helping behaviour of subordinates. more effort with removing experimentally added sand from the breeding shelter [42]. Experimental groups comprised of Materials and Methods 4.3362.19 subordinates .15 mm SL (mean 6 SD; range 2–8 subordinates/group, 16.0–48.5 mm SL). On average, 1.7361.58 Ethics statement large subordinates (.37.5 mm SL), 1.7361.16 small subordinates This study made use of large cages (see below) that were (25–37.5 mm SL) and 0.8761.41 juveniles (15–24.5 mm SL) were accepted by all N. pulcher enclosed. The fish showed no signs of present per group. A cage (26262 m; aluminium frame covered stress, and food intake rates were similar to N. pulcher outside the with sturdy plastic net, mesh size 2.562.5 mm to allow free cages (200–400 plankton bites per 15 min) and to previously plankton flow) was placed over the selected units and all piscivores reported data [40–42]. At least every four days, all fish in the cages were removed [41]. Cages were used to allow allocation of were monitored for signs of stress. This experiment was approved parentage of all potential candidates, which proved to be difficult by the Zambian Ministry of Agriculture, Food and Fisheries and it otherwise [55,56]. In total 39 such units were created and 78 complies with present laws of Zambia, the country where the study groups were enclosed for periods between 14 and 20 days. Before was performed. The study was approved by the Swiss Federal the quantitative recordings started, one to four subordinates per Veterinary Office Bern (licence no. 40/05). group were caught, sexed, measured (SL in mm, accuracy 0.5 mm) and marked by carefully excising half of a single fin ray Study species of the dorsal fin to facilitate identification [42]. Neolamprologus pulcher is a monomorphic cichlid When there were free swimming fry at the end of the two week’s occurring all around the sublittoral zone of the shores of Lake observation periods, all fish larger than 15 mm SL present in the Tanganyika [43]. The fish were studied by SCUBA diving cage were caught with hand nets and transparent Plexiglas tubes to between 8–11 m depth at Kasakalawe point, Zambia (8u46.8499S, be sexed, measured and fin-clipped [41]. All fish larger than 31u04.8829E) from September to November 2005 and 2006. 30 mm (SL) were sexed by close inspection of the genital papilla. Individuals live in social groups consisting of a dominant breeding After removing the stone covering the breeding shelter, fry were pair and usually 1–15 subordinates of both sexes that perform caught with help of the anaesthetic eugenol (1 part eugenol brood care, territory defence and maintenance [32,33]. Detailed dissolved in 4 parts 70% ethanol; [57]) and sampled wholly in descriptions of the behaviour have been provided elsewhere Eppendorf vials together with all fish #15 mm SL. In one case [33,44]. Groups contain on average 5 subordinate individuals during fry sampling, eggs were found and collected as well using .20 mm standard length (SL) [45] and the fish reach maturity at tweezers. Above water, eggs, fry, fish #15 mm SL and the fin clips about 30–35 mm SL [46]. Dominance among group members is of larger fish were stored in 95% ethanol for future DNA analyses. determined by size differences, even if small [34]. Large group members feed predominantly on zooplankton in the water column Genotyping [32,47], whereas small immature individuals also feed on benthic Ten polymorphic microsatellite loci were used to determine invertebrates within their territory [41]. In our study population, parentage of all broods (see Text S1 for details on loci and

PLoS ONE | www.plosone.org 2 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation microsatellite DNA markers). The software CERVUS3.0 [58] was dominant female, male or midpoint pair) predicted subordinate used to assign offspring based on exclusion. When offspring could reproductive participation (Logistic Regression) and (2) whether not be assigned to a known male, the minimum number of sires relatedness (covariate) caused any effects of subordinate repro- was estimated using the program GERUD2.0 [59]; see Text S1 for ductive participation on helping behaviour in the above GLMMs. details). Pairwise relatedness (r) estimations between mature Alpha was set to 0.05 throughout and all data were tested two- subordinates (.30 mm SL) and dominants were calculated with tailed. All statistical analyses were performed with SPSS software the program KINGROUP v2_090501 [60]. We calculated (version 17.0, SPSS Inc., Chicago, IL, USA). relatedness between: mature subordinates and dominant males, mature subordinates and dominant females, and mature subordi- Results nates and the midpoint of the dominant pair (average r to the dominant pair) using the KINSHIP estimator [61] and back- In total 27 out of 78 groups (34.6%) produced fry during the ground allele frequencies calculated according to Konovalov & experimental period. However, due to practical reasons broods of Heg (2008) [62]. only 15 groups could be collected. From these 15 groups we collected 18 broods, rendering on average 16.469.1 fry and eggs Behavioural observations per brood (range 4–33). From the 295 offspring analysed (258 fry One large and one small subordinate were observed per group and 37 eggs) we determined both parents of 276 offspring, one in random order three times for 10 min each, using a PVC-plate, parent of 17 offspring, and no parent of one fry and one egg each soft pencil and a waterproof stopwatch. Observations were (Table 1; Table S1). The majority of offspring (88.8%) were performed between 08:30 and 16:45 h and all behaviours were assigned to the dominant breeding pair; dominant females sired recorded in frequencies of occurrence, except for the time spent 99.7% of all offspring and dominant males sired 88.8%. In six out inside the breeding shelter. Once every minute the focal of 18 broods (33.3%) the dominant male shared paternity with subordinates’ height in the water column and its distance from other males and in five out of these six cases the extra-pair sires the breeding shelter were estimated. Recorded behaviours were assigned to subordinates of the same group (27.8% of all included overt attacks, restrained aggressive displays, submissive clutches; size range of subordinate males siring offspring: 31– behaviour and territory maintenance [33]. 41 mm SL). If male subordinates participated in reproduction, they gained on average 11.8% paternity. One fry was assigned to a dominant male from a neighbouring group, accounting for 7.7% Experimental sand addition and egg predator exposure extra-pair offspring in this brood. Taken together, in the six Every group was exposed twice to two experimental manipu- clutches with extra-pair paternity, on average 11.1% of the young lations to create standardised estimates of helping propensity. In in the brood were not sired by the dominant male (range: 6.3– the sand addition trials, the breeding shelter was carefully half– 22.7%). Subordinate females had never participated in reproduc- covered with sand to induce digging behaviour, and digging tion. Seventeen young collected in one territory belonged to two frequencies of all group members were recorded for 10 min different size classes indicating two separate broods; all four larger [41,42,53]. The 10 min recording of digging behaviour started fry had been produced by the dominant female of a neighbouring after the first individual of the group began to dig, or after 5 min territory and three out of these four young were fathered by a male when no digging was shown until then. In the egg predator not included in the cage population, suggesting a recent territory exposure trials one or four Telmatochromis vittatus were presented for take-over by the current dominant pair preceding the experimen- 10 min in a clear Plexiglas presentation tube (length 15 cm, tal period. diameter 8.2 cm) at 5 cm distance from the breeding shelter Relatedness between mature subordinates and dominants was entrance [41,42]. The number of presented egg predators was low (Table 2) and comparable to previously reported data from increased from one in 2005 to four in 2006 to ascertain egg this study population [28,31]. No difference in relatedness (r) was predator movement during presentations. We recorded all aggression against the presented T. vittatus from all group members, and the activity of the presented fish. For details about Table 1. Offspring sired by different males. egg predator sizes and their activity levels see Text S1.

Statistical analyses Assigned fathers Offspring number Normality of distributions was analysed with the one-sample Father Kolmogorov-Smirnov test and all data were tested for homoge- Broods without extra-pair offspring (n = 12 broods) neity of variance. Means of the three observations per focal Dominant male 161 subordinate were calculated and if necessary, data were trans- formed using square root transformations. Normally distributed Broods with extra-pair paternity (n = 6 broods) data were analysed with independent samples t-tests, whereas non- Male group members (n = 5 broods) normally distributed data were analysed with Mann-Whitney Dominant male 100 U-tests. Large subordinate 14a The sand addition and egg predator exposure trials were Small subordinate 1 analysed using Generalized Linear Mixed Models (GLMM), with the occurrence of participation in reproduction as a fixed effect Other male (n = 4 broods) and year as a random effect. All GLMMs allowed for unequal Dominant non-group male 1 variances by adjusting the scaling parameter (deviance method Unknown male 18b [63]). Due to the small number of subordinates siring offspring, additional potential random effects like cage identity and date Note that one egg did not amplify, thus no parentage could be assigned. aLarge male subordinates sired offspring in four broods. were not taken into account to avoid loss of predictive power. To bIncluding 14 eggs of two broods collected at a breeding shelter with two fry assess whether relatedness might have affected these results, we cohorts produced by two different females. analysed (1) whether relatedness (continuous factor: subordinate to doi:10.1371/journal.pone.0025673.t001

PLoS ONE | www.plosone.org 3 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation

Table 2. Pairwise relatedness (mean r 6 SE) of mature subordinate males, subordinates females and all subordinates combined with dominant females, dominant males and midpoint dominant pair (average r between the dominant pair).

Midpoint r Dominant males Dominant females

Male subordinates (n = 13) 0.17260.033 0.09660.084 0.19960.055 Female subordinates (n = 16) 0.07060.067 0.01860.084 0.12260.081 Subordinates combined (n = 39)* 0.09660.242 0.05860.311 0.11660.296

*Includes n = 10 subordinates with unclear sex. All subordinates used to calculate relatedness were larger than 30 mm SL (mean SL: 38.765.4 mm; range: 30.0–48.5 mm SL). doi:10.1371/journal.pone.0025673.t002 detected between the dominant pair (midpoint r), the dominant again the random effect of year was corrected for, which this time male and the dominant female relatedness with subordinates siring was significant, as subordinates were digging more in 2006 than in part of the offspring versus subordinates that did not participate in 2005 (Wald x2 = 17.407, df = 1, p,0.001). Furthermore, these reproduction (logistic regression, n = 13: midpoint r vs. siring effect: results did not change when relatedness was added as a covariate Wald x2 = 0.699, p = 0.403; dominant male r vs. siring effect: Wald to the two GLMMs above. No effects of subordinate relatedness x2 = 0.729, p = 0.393; dominant female r vs. siring effect: Wald were detected for midpoint r, dominant male r and dominant x2 = 0.514, p = 0.473). female r, respectively, on defence effort against experimentally In the egg predator exposure trials, subordinates who sired part presented egg predators (0.394#p#0.922) and on digging effort in of the offspring showed more defence effort against experimentally the sand exposure trials (0.233#p#0.874). presented egg predators than same-sized subordinates that had not Finally, male subordinates that participated in reproduction participated in reproduction (GLMM, n = 15: siring effect: Wald tended to stay closer to the breeding shelter than non-participating 2 x = 6.181, degrees of freedom [df] = 1, p = 0.013; Fig. 1); and the subordinates (t-test: t13 = 1.857, p = 0.086), whereas no differences random effect of year was corrected for, but it was not significant were found in the other behaviours tested (Table S2). (Wald x2 = 2.691, df = 1, p = 0.101). In contrast, in the sand exposure trials no difference was found between the frequency of Discussion digging between subordinates siring offspring and those that did Our data suggest that male subordinates of N. pulcher participate not (GLMM: siring effect: Wald x2 = 0.110, df = 1, p = 0.740); and in reproduction in the field and, if successful, they apparently raise their brood care effort accordingly. Furthermore, relatedness between subordinates and dominants did not affect the likelihood of subordinates’ siring offspring. In addition, helping behaviour did not depend on relatedness between subordinates and dominants. Taken together, we can exclude kin selection as a factor explaining both subordinate parentage and helping behaviour, and the effect of subordinate parentage on their brood care effort remained significant after correcting for the (non- significant) effect of kinship. This indicates that current direct fitness benefits, such as the production of own offspring, are important for the performance and intensity of specific cooperative behaviours in subordinates of cooperatively breeding fish. In species with low relatedness between dominants and subordinates, receiving indirect fitness benefits through helping to raise offspring of relatives is improbable. This might make attempts to obtain direct fitness benefits via parentage more rewarding. In a few cooperatively breeding birds and mammals, subordinates partic- ipating in reproduction also increased their helping effort in the field [22,29,64,65] and a laboratory study of N. pulcher revealed that female subordinates may increase alloparental care in response to their participation in reproduction [30,38]. These data suggest that more generally, directs fitness benefits might be an important modifier of subordinate helping intensities in cooperative breeders. Theoretical arguments suggest that in groups with multiple males, the fitness of dominant males may increase when subordinates sire only small parts of the offspring because of conflict reduction [66]. This and the increased levels of subordinate helping could select for reproductive concessions Figure 1. Per capita frequency of defence per 10 min for subordinates who participated in reproduction (black circle) provided by dominants [67]. Previous results suggest that and non-participating subordinates (open circle) against participation of reproduction of male [37,38] and female presented egg predators. Means 6 SE are shown; * p,0.05. subordinates in N. pulcher [68] is compatible with tug-of-war doi:10.1371/journal.pone.0025673.g001 models of reproductive skew. Heuristic skew models, however,

PLoS ONE | www.plosone.org 4 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation provide little predictive power to explain reproductive skew in subordinate males are likely to participate in reproduction [17], cooperatively breeding fish groups due to the complexity of the and testis mass correlates positively between dominant male mechanisms involved [69]. In N. pulcher it is unlikely that breeders and their male subordinates, which suggests adjustments dominant males are in full control of reproduction, because they to the level of intragroup sperm competition [74]. Our results frantically attempt to prevent subordinates participating in show that in N. pulcher, relatedness did not differ between the spawning [34], and more dominant reproduction is lost to dominant pair and subordinates that sired part of the offspring and subordinates if multiple subordinate males are present in the those that did not. Furthermore, when testing for effects of group [38,70]. Furthermore, dominant males show more subordinate parentage and relatedness to dominants on subordi- aggression towards male than female subordinates [71], especially nate defence against egg predators, only subordinate parentage during reproductive periods [70]. This implies that male showed significant effects, but not relatedness levels between subordinates entail costs to dominant males mainly by parasit- subordinates and dominants. This indicates that relatedness does izing reproduction (cf. [35]). In compensation for these fitness not strongly affect subordinate helping effort in N. pulcher,orat costs, dominant males might benefit from increased brood care least not as strongly as subordinate parentage does. levels provided by male subordinates that have shared in Our results show that male subordinates sharing in reproduc- reproduction. As dominant males provide virtually no help in tion tended to stay closer to the breeding shelter, which might brood care, apart from deterring large piscivores [33,41,42,53], serve as a guarding function. Previously we have reported size- they may benefit more via increased male subordinate aid than dependent sharing of tasks among subordinates of this species, what they lose by sharing part of reproduction. In general, the with small subordinates specialising in defence against egg costs caused by subordinate group members have been predicted predators [42]. Our new findings suggest that in addition to size to be partly or fully compensated for by their cooperative effort if dependence, the effort of male subordinates in brood care and helpers pay to stay, but helping should not provide net benefits to protection may also depend on their participation in reproduction. dominants [72]. In other words, the rent helpers pay to be In banded mongooses, especially male subordinates contribute allowed to stay in the territory (cf. [36,40,42,51–53,73]) merely more to guarding during times of high energy expenditure and the serves as cost compensation. This predicts that the higher the survival rates of young increases with the number of guards [75]. costs caused by subordinates, the more they should help, which However, in contrast to N. pulcher, yearling non- has been supported by our data. Another incentive to increase breeding males tended to make higher individual contributions to brood care levels when successfully sharing in reproduction is the the care of pups than yearling breeding males that may have fact that some of the young benefitting from care will be own successfully participated in offspring production [65]. offspring. Currently, we cannot differentiate between these two Digging behaviour seems to be less flexible than defence in N. potential functional causes of the positive correlation between the pulcher, as the removal of experimentally added sand from the reproductive participation of helpers and their brood care effort. breeding shelter was not related to the subordinates’ share in In previous studies, extra-group paternity [55,56] and reproduction. This confirms previous results of similar sand subordinate maternity [56] had been observed under natural addition trials revealing little plasticity of large subordinates in conditions in N. pulcher, however reproductive participation of response to varying digging demands. For instance, no difference male subordinates was not detected. Our results confirm the in digging intensity was found among large subordinates between: (i) levels of male subordinate reproductive participation found in low and high neighbour densities simulating variation in space laboratory experiments [35,37,38]. This rather moderate repro- competition [52]; (ii) natural and reduced food conditions [41]; (iii) ductive participation of subordinates is difficult to detect when isolated and group living individuals [76]; (iv) low and high sample sizes per brood are small (in the previously published densities of egg predators [42]; and (v) differences in reproductive studies [55,56], the mean number of sampled offspring per brood status of groups, i.e. with or without free-swimming fry (Bruintjes had been 3.6 and 3.9, respectively). Additionally, there is R, Louter M & Taborsky M, unpubl. data). evidence that in previous studies the paternity of male In conclusion, our data show that in cooperatively breeding subordinates may have remained undetected because of their cichlids male subordinates can gain parentage in the field, and that eviction or dispersal before parentage could be determined [55]. this might affect their effort spent on specific helping behaviours. By using large underwater cages we prevented this occurring in Our results stress that current direct fitness benefits (participation our study and therefore obtained genetic samples of almost all in reproduction) might be of importance in modifying subordinate potential reproductive individuals present during offspring helping effort in cooperative breeders. production. Nevertheless, one egg and one fry (0.7% of all offspring) could not be assigned to any parent and in 17 out of Supporting Information 295 offspring (5.8% of all offspring); one parent could not be assigned. These latter cases might have resulted at least partly Table S1 Parentage of 18 broods collected from 15 from offspring production shortly before the experiment started groups in the field. (see Results section). Furthermore, it should be considered that (DOCX) we incorporated only groups with rather low numbers of Table S2 Behavioural comparisons of subordinates subordinates in this study, which might result in an underesti- with and without parentage. The table shows all focal mation of male subordinate parentage due to a potential behaviours which were tested with independent sample t-tests, exponential increase of subordinate paternity with increasing except for submissiveness, which was tested with a Mann-Whitney numbers of male subordinates in the group [36,62]. U-test. 0.05,p-values,0.10 are underlined. Reproductive participation of male subordinates in N. pulcher (DOCX) had been assumed to be ‘unlikely’ due to their relatively low investment in testis and sperm quality [17]. It remains to be tested Text S1 This supplementary text contains additional if subordinates participating in reproduction show a higher details on genotyping, parentage assignment and egg investment in testis and sperm quality compared to the ones that predator presentation trials. did not. For instance, in the cichlid Julidochromis ornatus all (DOCX)

PLoS ONE | www.plosone.org 5 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation

Acknowledgments this paper to the late Renske Hekman in gratitude for her friendship and collaboration. We thank Dr. H. Phiri, R. Shapola and L. Makasa from the Department of Fisheries, Mpulungu, Zambia for their continuous support of our project Author Contributions and the Swiss National Science Foundation for financial support. Moreover, we are grateful to Renske Hekman and Marina Louter for Conceived and designed the experiments: RB DH MT. Performed the help gathering the data, Kate Martin for proof reading and to three experiments: RB. Analyzed the data: RB DH. Contributed reagents/ anonymous referees for helpful comments on the manuscript. We dedicate materials/analysis tools: DB. Wrote the paper: RB DH MT. Performed all laboratory work: DB.

References 1. Stacey PB, Koenig WD (1990) Cooperative breeding in birds: Long-term studies 27. Wright J, McDonald PG, te Marvelde L, Kazem AJN, Bishop CM (2010) of ecology and behavior. Cambridge: Cambridge University Press. Helping effort increases with relatedness in bell miners, but ‘unrelated’ helpers of 2. Taborsky M (1994) Sneakers, satellites and helpers: parasitic and cooperative both sexes still provide substantial care. Proceedings of the Royal Society B- behavior in fish reproduction. Advances in the Study of Behavior 23: 1–100. Biological Sciences 277: 437–445. 3. Cockburn A (1998) Evolution of helping behavior in cooperatively breeding 28. Stiver KA, Dierkes P, Taborsky M, Gibbs HL, Balshine S (2005) Relatedness birds. Annual Review of Ecology and Systematics 29: 141–177. and helping in fish: examining the theoretical predictions. Proceedings of the 4. Cochran GR, Solomon NG (2000) Effects of food supplementation on the social Royal Society B-Biological Sciences 272: 1593–1599. organization of prairie voles (Microtus ochrogaster). Journal of Mammalogy 81: 29. Canestrari D, Marcos JM, Baglione V (2005) Effect of parentage and relatedness 746–757. on the individual contribution to cooperative chick care in carrion crows Corvus 5. Heg D, Bachar Z (2006) Cooperative breeding in the lake tanganyika cichlid corone corone. Behavioral Ecology and Sociobiology 57: 422–428. Julidochromis ornatus. Environmental Biology of Fishes 76: 265–281. 30. Heg D, Jutzeler E, Mitchell JS, Hamilton IM (2009) Helpful Female 6. Hamilton WD (1963) The evolution of altruistic behavior. American Naturalist Subordinate Cichlids Are More Likely to Reproduce. Plos One 4. 97: 354–356. 31. Dierkes P, Heg D, Skubic E, Taborsky M, Achmann R (2005) Genetic 7. Hamilton WD (1964) The genetical evolution of social behaviour I & II. Journal relatedness in groups is sex-specific and declines with age of helpers in a of Theoretical Biology 7: 1–52. cooperatively breeding cichlid. Ecology Letters 8: 968–975. 8. Lehmann L, Keller L (2006) The evolution of cooperation and altruism - a 32. Taborsky M, Limberger D (1981) Helpers in fish. Behavioral Ecology and general framework and a classification of models. Journal of Evolutionary Sociobiology 8: 143–145. Biology 19: 1365–1376. 33. Taborsky M (1984) Broodcare helpers in the cichlid fish Lamprologus brichardi: 9. Zahavi A (1995) Altruism as a handicap - the limitations of kin selection and their costs and benefits. Behaviour 32: 1236–1252. reciprocity. Journal of Avian Biology 26: 1–3. 34. Heg D, Bachar Z, Brouwer L, Taborsky M (2004) Predation risk is an ecological 10. Kokko H, Johnstone RA, Clutton-Brock TH (2001) The evolution of constraint for helper dispersal in a cooperatively breeding cichlid. Proceedings of cooperative breeding through group augmentation. Proceedings of the Royal the Royal Society of London Series B-Biological Sciences 271: 2367–2374. Society of London Series B-Biological Sciences 268: 187–196. 35. Dierkes P, Taborsky M, Kohler U (1999) Reproductive parasitism of broodcare 11. Gaston AJ (1978) Evolution of group territorial behavior and cooperative helpers in a cooperatively breeding fish. Behavioral Ecology 10: 510–515. breeding. American Naturalist 112: 1091–1100. 36. Taborsky M (1985) Breeder-helper conflict in a cichlid fish with broodcare 12. Kokko H, Johnstone RA, Wright J (2002) The evolution of parental and helpers: an experimental analysis. Behaviour 95: 45–75. alloparental effort in cooperatively breeding groups: when should helpers pay to 37. Heg D, Bergmu¨ller R, Bonfils D, Otti O, Bachar Z, et al. (2006) Cichlids do not stay? Behavioral Ecology 13: 291–300. adjust reproductive skew to the availability of independent breeding options. 13. Magrath RD, Whittingham LA (1997) Subordinate males are more likely to help Behavioral Ecology 17: 419–429. if unrelated to the breeding female in cooperatively breeding white-browed 38. Heg D, Jutzeler E, Bonfils D, Mitchell JS (2008) Group composition affects male scrubwrens. Behavioral Ecology and Sociobiology 41: 185–192. reproductive partitioning in a cooperatively breeding cichlid. Molecular Ecology 14. Dickinson JL (2004) A test of the importance of direct and indirect fitness 17: 4359–4370. benefits for helping decisions in western bluebirds. Behavioral Ecology 15: 39. Wong M, Balshine S (2011) The evolution of cooperative breeding in the African 233–238. cichlid fish, Neolamprologus pulcher. Biological Reviews. 15. Skubic E, Taborsky M, McNamara JM, Houston AI (2004) When to parasitize? 40. Balshine-Earn S, Neat FC, Reid H, Taborsky M (1998) Paying to stay or paying A dynamic optimization model of reproductive strategies in a cooperative to breed? Field evidence for direct benefits of helping behavior in a cooperatively breeder. Journal of Theoretical Biology 227: 487–501. breeding fish. Behavioral Ecology 9: 432–438. 16. Bergmuller R, Johnstone RA, Russell AF, Bshary R (2007) Integrating 41. Bruintjes R, Hekman R, Taborsky M (2010) Experimental global food reduction cooperative breeding into theoretical concepts of cooperation. Behavioural raises resource acquisition costs of brood care helpers and reduces their helping Processes 76: 61–72. effort. Functional Ecology 24: 1054–1063. 17. Fitzpatrick JL, Desjardins JK, Stiver KA, Montgomerie R, Balshine S (2006) Male reproductive suppression in the cooperatively breeding fish Neolamprologus 42. Bruintjes R, Taborsky M (2011) Size dependent task specialization in a pulcher. Behavioral Ecology. pp 25–33. cooperative cichlid in response to experimental variation of demand. Animal 18. Awata S, Munehara H, Kohda M (2005) Social system and reproduction of Behaviour 81: 387–394. helpers in a cooperatively breeding cichlid fish (Julidochromis ornatus) in Lake 43. Duftner N, Sefc KM, Koblmuller S, Salzburger W, Taborsky M, et al. (2007) Tanganyika: field observations and parentage analyses. Behavioral Ecology and Parallel evolution of facial stripe patterns in the /pulcher Sociobiology 58: 506–516. species complex endemic to Lake Tanganyika. Molecular Phylogenetics and 19. Whittingham LA, Dunn PO, Magrath RD (1997) Relatedness, polyandry and Evolution 45: 706–715. extra-group paternity in the cooperatively-breeding white-browed scrubwren 44. Heg D, Brouwer L, Bachar Z, Taborsky M (2005) Large group size yields group (Sericornis frontalis). Behavioral Ecology and Sociobiology 40: 261–270. stability in the cooperatively breeding cichlid Neolamprologus pulcher. Behaviour 20. Goossens B, Graziani L, Waits LP, Farand E, Magnolon S, et al. (1998) Extra- 142: 1615–1641. pair paternity in the monogamous Alpine marmot revealed by nuclear DNA 45. Balshine S, Leach B, Neat F, Reid H, Taborsky M, et al. (2001) Correlates of microsatellite analysis. Behavioral Ecology and Sociobiology 43: 281–288. group size in a cooperatively breeding cichlid fish (Neolamprologus pulcher). 21. Davies NB (1992) Dunnock Behaviour and Social Evolution. Oxford: Oxford Behavioral Ecology and Sociobiology 50: 134–140. University Press. 46. Heg D, Rothenberger S, Schu¨rch R (2011) Habitat saturation, benefits of 22. Hartley IR, Davies NB, Hatchwell BJ, Desrochers A, Nebel D, et al. (1995) The philopatry, relatedness, and the extent of co-operative breeding in a cichlid. Polygynandrous Mating System of the Alpine Accentor, Prunella-Collaris. II. Behavioral Ecology 22: 82–92. Multiple Paternity and Parental Effort. Animal Behaviour 49: 789–803. 47. Gashagaza MM, Nagoshi M (1986) Comparative study on the food habits of six 23. Nam KB, Simeoni M, Sharp SP, Hatchwell BJ (2010) Kinship affects investment species of Lamprologus (Osteichthyes: Cichlidae). African Study Monographs 6: by helpers in a cooperatively breeding bird. Proceedings of the Royal Society B- 37–44. Biological Sciences 277: 3299–3306. 48. Heg D, Heg-Bachar Z, Brouwer L, Taborsky M (2008) Experimentally induced 24. Kingma SA, Hall ML, Peters A (2011) Multiple Benefits Drive Helping Behavior helper dispersal in colonially breeding cooperative cichlids. Environmental in a Cooperatively Breeding Bird: An Integrated Analysis. American Naturalist Biology of Fishes 83: 191–206. 177: 486–495. 49. Brouwer L, Heg D, Taborsky M (2005) Experimental evidence for helper effects 25. Armitage KB, Schwartz OA (2000) Social enhancement of fitness in yellow- in a cooperatively breeding cichlid. Behavioral Ecology 16: 667–673. bellied marmots. Proceedings of the National Academy of Sciences of the United 50. Taborsky B, Skubic E, Bruintjes R (2007) Mothers adjust egg size to helper States of America 97: 12149–12152. number in a cooperatively breeding cichlid. Behavioral Ecology 7: 652–657. 26. Le Vin AL, Mable BK, Taborsky M, Heg D, Arnold KE (2011) Individual 51. Bergmu¨ller R, Taborsky M (2005) Experimental manipulation of helping in a variation in helping in a cooperative breeder: relatedness versus behavioural cooperative breeder: helpers ‘pay to stay’ by pre-emptive appeasement. Animal type. Animal Behaviour: In press. Behaviour 69: 19–28.

PLoS ONE | www.plosone.org 6 October 2011 | Volume 6 | Issue 10 | e25673 Paternity Affects Cooperation

52. Bruintjes R, Taborsky M (2008) Helpers in a cooperative breeder pay a high 65. Gilchrist JS, Russell AF (2007) Who cares? Individual contributions to pup care price to stay: effects of demand, helper size and sex. Animal Behaviour 75: by breeders vs non-breeders in the cooperatively breeding banded mongoose 1843–1850. (Mungos mungo). Behavioral Ecology and Sociobiology 61: 1053–1060. 53. Heg D, Taborsky M (2010) Helper Response to Experimentally Manipulated 66. Houston AI, Gasson CE, McNamara JM (1997) Female choice of matings to Predation Risk in the Cooperatively Breeding Cichlid Neolamprologus pulcher. Plos maximize parental care. Proceedings of the Royal Society of London Series B- One 5. Biological Sciences 264: 173–179. 54. Le Vin AL, Mable BK, Arnold KE (2010) Kin recognition via phenotype 67. Keller L, Reeve HK (1994) Partitioning of Reproduction in Animal Societies. matching in a cooperatively breeding cichlid, Neolamprologus pulcher. Animal Trends in Ecology & Evolution 9: 98–102. Behaviour 79: 1109–1114. 68. Heg D, Hamilton IM (2008) Tug-of-war over reproduction in a cooperatively 55. Dierkes P, Taborsky M, Achmann R (2008) Multiple paternity in the breeding cichlid. Behavioral Ecology and Sociobiology 62: 1249–1257. cooperatively breeding fish Neolamprologus pulcher. Behavioral Ecology and 69. Taborsky M (2009) Reproductive skew in cooperative fish groups: virtue and Sociobiology 62: 1581–1589. limitations of alternative modeling approaches. In: Hager R, Jones CB, eds. 56. Stiver KA, Fitzpatrick JL, Desjardins JK, Balshine S (2009) Mixed parentage in Reproductive Skew in Vertebrates: Proximate and Ulitimate Causes. Cam- Neolamprologus pulcher groups. Journal of Fish Biology 74: 1129–1135. bridge: Cambrige University Press. pp 265–304. 57. Kreiberg H (2000) Stress and anesthesia. In: Ostrander GK, ed. The Laboratory 70. Mitchell JS, Jutzeler E, Heg D, Taborsky M (2009) Gender Differences in the Fish. New York: Academic Press. pp 503–511. Costs that Subordinate Group Members Impose on Dominant Males in a 58. Kalinowski ST, Taper ML, Marshall TC (2007) Revising how the computer Cooperative Breeder. Ethology 115: 1162–1174. program CERVUS accommodates genotyping error increases success in 71. Mitchell JS, Jutzeler E, Heg D, Taborsky M (2009) Dominant members of paternity assignment. Molecular Ecology 16: 1099–1106. 59. Jones AG (2005) GERUD 2.0: a computer program for the reconstruction of cooperatively-breeding groups adjust their behaviour in response to the sexes of parental genotypes from half-sib progeny arrays with known or unknown their subordinates. Behaviour 146: 1665–1686. parents. Molecular Ecology Notes 5: 708–711. 72. Hamilton IM, Taborsky M (2005) Unrelated helpers will not fully compensate 60. Konovalov DA, Manning C, Henshaw MT (2004) KINGROUP: a program for for costs imposed on breeders when they pay to stay. Proceedings of the Royal pedigree relationship reconstruction and kin group assignments using genetic Society B-Biological Sciences 272: 445–454. markers. Molecular Ecology Notes 4: 779–782. 73. Bergmu¨ller R, Heg D, Taborsky M (2005) Helpers in a cooperatively breeding 61. Queller DC, Goodnight KF (1989) Estimating relatedness using genetic markers. cichlid stay and pay or disperse and breed, depending on ecological constraints. Evolution 43: 258–275. Proceedings of the Royal Society B-Biological Sciences 272: 325–331. 62. Konovalov DA, Heg D (2008) A maximum-likelihood relatedness estimator 74. Awata S, Heg D, Munehara H, Kohda M (2006) Testis size depends on social allowing for negative relatedness values. Molecular Ecology Resources 8: status and the presence of male helpers in the cooperatively breeding cichlid 256–263. Julidochromis ornatus. Behavioral Ecology. pp 1–8. 63. Norusis MJ (2010) PASW Statistics 18 Advanced Statistical Procedures 75. Cant MA (2003) Patterns of helping effort in co-operatively breeding banded Companion. Upper Saddle River: Prentice Hall. mongooses (Mungos mungo). Journal of Zoology 259: 115–121. 64. Whittingham LA, Dunn PO (1998) Male parental effort and paternity in a 76. Schu¨rch R, Heg D (2010) Life history and behavioral type in the highly social variable mating system. Animal Behaviour 55: 629–640. cichlid Neolamprologus pulcher. Behavioral Ecology 21: 588–598.

PLoS ONE | www.plosone.org 7 October 2011 | Volume 6 | Issue 10 | e25673